Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 7 Jul 1997 07:35:43 -0400 (EDT)
From:      Thomas David Rivers <ponds!rivers@dg-rtp.dg.com>
To:        ponds!lakes.water.net!rivers, ponds!sdf.com!tom
Cc:        ponds!FreeBSD.ORG!freebsd-hackers
Subject:   Re: Why news expiration is sooo slowww with 2.2.x.
Message-ID:  <199707071135.HAA01419@lakes.water.net>

next in thread | raw e-mail | index | archive | help
> 
> On Sun, 6 Jul 1997, Thomas David Rivers wrote:
> 
> (BTW, can you stop your mail headers from getting mangled?)

 I've been working on it for some time.  Often times, it's
an incorrectly configured host between my machine and someone
else.   Can you send me the header you receive from this?

> 
> > If you recall, I mentioned that my news expiration took
> > a serious turn for the worse when I installed 2.2.1.  That is,
> > expirations went from a few hours to 2-3 days (sometimes even
> > longer now...)
> 
>   I use a 2.2.1 on a news server, although I should upgrade...
> 
> > I believe I've tripped over the issue.  It would appear that
> > readdir et. al. is taking a *long* time.  If I go to a directory
> > which has a lot of files and simply do an 'ls', it can
> > take on the order of 20-30 minutes...  I wouldn't expect the sort
> > to be the culprit here...  and, since I'm not asking for any
> > file information, I think we can eliminate stat() as a potential
> > culprit.
> 
>   What options are you using on the "ls"?  "-l" causes lots of getpwuid()
> calls.  "-f" will avoid sorting.  How many files were actually in that
> directory?  If the directory is screwed, and you have a million files in
> it, 20-30 minutes is probably normal, and chances are the expiry will
> never remove them either.

  No -l, just 'ls' (to avoid the stat()) - the directory isn't messed
up - an fsck indicates everything is just fine.  It's just that there
are a lot of files.

  A 
	/bin/ls -1 | wc -l

reveals:

> 
> > I thought I would try to investigate what the problem was, but, 
> > before that; I'd send out the standard "has anyone seen/done anything
> > in this area already?"   Has anyone else noticed an issue with
> > readdir()?
> 
>   No.  I have over 20GB of news spool space.  It only takes a couple of
> hours to expire.  Uptime is currently at 37 days, running 2.2.1

 I should have mentioned, this is a 386-33 (with an Intel 387) and
only 8megs of RAM.  Prior to 2.2.1, it took only a few hours to expire; 
now it takes many days....   Examining swapinfo during this process
indicates swap is mostly free, so I don't believe my constrained memory
is the problem (I'm not even swapping...)   Certainly my CPU is
slow, but it wasn't any faster prior to version 2.2.1.. :-)

	- Thanks -
	- Dave R. -





Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199707071135.HAA01419>